"PeterPan" <<zenothing[at]hotmail.com>> wrote:
> But with Pike v7.8 release 116 the result is:
>
> ([ /* 1 element */
> "\u00e5\u00a4\u00a7\u00e5\u00ae\u00b6\u00e5\u00a5\u00bd": 1
> ])
>
> Why escape it ?
Pike has only switched escape method from \oooooo to \uxxxx. If you
update your VARINFO class to handle \u escapes then it should work
again.
That aside, the stdout and stderr handling could be better when they
are connected to a terminal. In that case pike should look at the
locale settings and encode charsets appropriately, so that the output
streams can accept unicode strings. I don't quite know how the
terminal stuff works, but it obviously doesn't handle this entirely
well.
|